Granted Patent B2

US 12,731,112 · App. 19/293,963 · Granted Sep 8, 2026

Inventors: Alan W. Hecht (Richmond, CA); Frank Henry Filippini (Palm Desert, CA); Matthew Evan Greber (Corte Madera, CA)

Assignee: Wells Fargo Bank, N.A.

G06Q20/042G06T7/0002G06T7/60G06T7/62G06T7/70G06V10/44G06T2207/30168

Abstract

Claims (75)

1 . A method comprising:

capturing, at a user device, a first video of a first side of each of a plurality of documents;

outputting a notification to the user device requesting each of the plurality of documents to be flipped to a second side;

capturing, at the user device, a second video of the second side of each of the plurality of documents;

extracting a first set of respective individual images of the first side of each document of the plurality of documents from the first video;

extracting a second set of respective individual images of the second side of each document of the plurality of documents from the second video;

comparing, using processing circuitry of the user device, geometrical features of the first side of each of the plurality of documents as represented in the first set of respective individual images to geometrical features of the second side of each of the plurality of documents as represented in the second set of respective individual images, the geometrical features including at least one of a size of a document, a fold, a tear, a stain, or a crease;

based on the comparison, selecting, for each image from the first set of respective individual images, a corresponding image from the second set of respective individual images to form a plurality of pairs of images, each pair of images of the plurality of pairs of images representing a single document of the plurality of documents; and

outputting, from the user device, the plurality of pairs of images.

2 . The method of claim 1 , further comprising:

receiving a user input identifying a number of documents of the plurality of documents;

determining that a number of captured documents in the first video or the second video does not match the number of documents of the plurality of documents; and

outputting a second notification to the user device based on the determination, the second notification requesting the first video or the second video to be recaptured.

3 . The method of claim 1 , further comprising:

determining that a number of documents captured in the first video or the second video does not match the number of documents in the plurality of documents; and

outputting a third notification to the user device based on the determination, the third notification requesting that at least one of the first video or the second video be recaptured.

4 . The method of claim 1 , wherein the first side is a front side of the plurality of documents and the second side is a back side of the plurality of documents.

5 . The method of claim 1 , further comprising:

providing instructions on a user interface, while capturing the first video or the second video, the instructions including an alert to move the user device in a particular direction.

6 . The method of claim 1 , further comprising:

determining that a number of the first set of respective individual images does not match a second number of the second set of respective individual images; and

outputting a fourth notification to the user device based on the determination, the fourth notification requesting that the first video or the second video be recaptured.

7 . The method of claim 1 , further comprising:

converting the plurality of pairs of images to a set of Tag Image File Format (TIFF) files; and

outputting, via an application programming interface (API), the set of TIFF files to a document depository.

8 . The method of claim 1 , wherein comparing the geometrical features of the first side of each of the plurality of documents as represented in the first set of respective individual images to the geometrical features of the second side of each of the plurality of documents as represented in the second set of respective individual images includes further comparing at least one of:

an orientation of a respective document determined from a magnetic ink character recognition (MICR) line; or

the orientation of the respective document determined from text orientation.

9 . The method of claim 1 , further comprising:

determining, after capturing the first video or the second video, that one or more documents of the plurality of documents are overlapping; and

outputting a fifth notification to the user device based on the determination, the fifth notification requesting that the first video or the second video be recaptured.

10 . The method of claim 1 , further comprising:

before outputting the notification to the user device requesting each of the plurality of documents to be flipped to the second side, determining that the first side of each document captured in the first video satisfies a quality metric; and

before extracting the second set of respective individual images of the second side of each document of the plurality of documents from the second video, determining that the second video of the second side of each of the plurality of documents satisfies the quality metric.

11 . The method of claim 10 , wherein determining that the first side of each document captured in the first video satisfies the quality metric includes determining that a quality score for the first side of each document in a frame of the first video exceeds a quality score threshold.

12 . The method of claim 1 , further comprising determining that a side of a captured document in a third video fails to satisfy a quality metric; and

in response, outputting a sixth notification to the user device based on the determination, the sixth notification requesting that the third video be recaptured.

13 . The method of claim 1 , further comprising:

determining that one or more images in the first set of respective individual images or in the second set of respective individual images are blurry; and

outputting a seventh notification to the user device based on the determination, the seventh notification requesting that the first video or the second video be recaptured.

14 . At least one non-transitory machine readable medium including instructions, which when executed by processing circuitry of a user device, cause the processing circuitry to perform operations to:

capture a first series of images of a first side of each of a plurality of documents;

determine that the first side of each document captured in the first series of images satisfies a quality metric;

output a notification to the user device based on the determination, the notification requesting each of the plurality of documents be flipped to a second side;

capture a second series of images of the second side of each of the plurality of documents;

determine that the second series of images of the second side of each of the plurality of documents satisfies the quality metric;

extract a first set of respective individual images of the first side of each document of the plurality of documents from the first series of images;

extract a second set of respective individual images of the second side of each document of the plurality of documents from the second series of images;

compare geometrical features of the first side of each of the plurality of documents as represented in the first set of respective individual images to geometrical features of the second side of each of the plurality of documents as represented in the second set of respective individual images, the geometrical features including at least one of a size of a document, a fold, a tear, a stain, or a crease;

based on the comparison, select, for each image from the first set of respective individual images, a corresponding image from the second set of respective individual images to form a plurality of pairs of images, each pair of images of the plurality of pairs of images representing a single document of the plurality of documents; and

output, from the user device, the plurality of pairs of images.

15 . The at least one non-transitory machine readable medium of claim 14 , wherein the instructions further cause the processing circuitry to perform operations to:

receive a user input identifying a number of documents of the plurality of documents;

determine that a number of captured documents in the first series of images or the second series of images does not match the number of documents of the plurality of documents; and

output a second notification to the user device based on the determination, the second notification requesting the first series of images or the second series of images to be recaptured.

16 . The at least one non-transitory machine readable medium of claim 14 , wherein the instructions further cause the processing circuitry to perform operations to:

display guidance on a user interface, while capturing the first series of images or the second series of images, the instructions including an alert to move the user device in a particular direction.

17 . The at least one non-transitory machine readable medium of claim 14 , wherein the instructions further cause the processing circuitry to perform operations to:

determine that a number of the first set of respective individual images does not match a second number of the second set of respective individual images; and

output a third notification to the user device based on the determination, the third notification requesting that the first series of images or the second series of images be recaptured.

18 . The at least one non-transitory machine readable medium of claim 14 , wherein the instructions further cause the processing circuitry to perform operations to:

convert the plurality of pairs of images to a set of Tag Image File Format (TIFF) files; and

output, via an application programming interface (API), the set of TIFF files to a document depository.

19 . At least one non-transitory machine readable medium including instructions, which when executed by processing circuitry of a user device, cause the processing circuitry to perform operations to:

capture a first video of a first side of each of a plurality of documents;

determine that the first side of each document captured in the first video satisfies a first quality metric;

output a notification to the user device based on the determination, the notification requesting each of the plurality of documents be flipped to a second side;

capture a second video of the second side of each document of the plurality of documents;

determine that the second video of the second side of each the plurality of documents satisfies a second quality metric;

extract a first set of respective individual images of the first side of each document of the plurality of documents from the first video;

extract a second set of respective individual images of the second side of each document of the plurality of documents from the second video;

compare geometrical features of the first side of each of the plurality of documents as represented in the first set of respective individual images to geometrical features of the second side of each of the plurality of documents as represented in the second set of respective individual images, the geometrical features including at least one of a size of a document, a fold, a tear, a stain, or a crease;

based on the comparison, select, for each image from the first set of respective individual images, a corresponding image from the second set of respective individual images to form a plurality of pairs of images, each pair of images of the plurality of pairs of images representing a single document of the plurality of documents; and

output, from the user device, the plurality of pairs of images.

20 . The at least one non-transitory machine readable medium of claim 19 , wherein the first quality metric applies to a front of each document and the second quality metric applies to a back of each document.

Assignments (1)

Continuity (2)

References Cited (16)